Environmental Sound Monitoring And System Implementation Based On Android Platform

With the acceleration of the construction of smart cities,social security has become a topic of eager attention by the public,and the role of security monitoring in the field of national defense and people’s livelihood has become increasingly prominent.Due to its ability to characterize the surrounding environment,environmental sound has become a hot research topic in recent years and has been successfully applied in areas such as security protection,urban noise,and transportation.The environmental sound monitoring system uses endpoint detection to capture environmental sound events,reduces the amount of system data transmission and improves the recognition rate of environmental sound events.However,in a low signal-to-noise ratio environment,traditional endpoint detection algorithms have problems such as low hit rate,high over-segmentation rate,and low accuracy due to the interference of strong noise.In addition,although related projects have made significant progress in environmental sound monitoring research,due to the fact that the commonly used monitoring methods are expensive to measure devices and equipment and require professional operation,the deployment cost is too high,and they are vulnerable to environmental conditions.It is difficult to realize large-scale environmental sound monitoring.Therefore,the following solutions are proposed for the above-mentioned problems.(1)Aiming at the problem of low performance of traditional endpoint detection algorithms in low signal-to-noise ratio environments,an improved energy-zero-product endpoint detection algorithm based on sonogram enhancement is proposed.The algorithm first uses the improved spectral subtraction method to denoise the signal to improve the signal-to-noise ratio;then the denoised sonogram is sequentially processed by erosion,binarization and dilatation to achieve the enhancement of the sonogram;finally,the energy-zero-product feature is extracted and use the double-threshold method for endpoint detection.Experimental results show that in a low signal-to-noise ratio environment of-10 d B to 5d B,the improved algorithm compares the three traditional algorithms of frequency band variance,energy-to-entropy ratio and cepstrum distance,the average hit rate is increased by up to 9.0%,and the average over-segmentation rate is reduced by up to 114.0%,and the average accuracy is increased by up to 38.8%.This shows that the improved endpoint detection algorithm proposed in this paper is more suitable for capturing environmental sound events in a low signal-to-noise ratio environment.(2)Aiming at the problem that commonly used environmental sound monitoring methods are susceptible to deployment costs and environmental restrictions and cannot achieve large-scale monitoring,an environmental sound monitoring system based on the Android platform is proposed.By deploying stationary monitoring nodes,using smart phones as mobile monitoring nodes,and adopting group intelligence sensing methods,the system realizes low-cost,large-scale environmental sound monitoring that combines mobile and stationary monitoring nodes.This paper applies the improved endpoint detection algorithm to the proposed system,and realizes the combination of theoretical research and practical application,and deploys the system in the actual campus environment.Through relevant experimental tests and data analysis,the system is verified feasibility.This shows that the system proposed in this paper is more suitable for environmental sound monitoring in a wide range of situations.